RESEARCH ARTICLE

NML-mediated rRNA base methylation links ribosomal subunitformation to cell proliferation in a p53-dependent mannerTsuyoshi Waku1,*, Yuka Nakajima2,3,‡, Wataru Yokoyama3, Naoto Nomura3, Koichiro Kako2,3, Akira Kobayashi4,Toshiyuki Shimizu1 and Akiyoshi Fukamizu2,3

ABSTRACTRibosomal RNAs (rRNAs) act as scaffolds and ribozymes inribosomes, and these functions are modulated by post-transcriptionalmodifications. However, the biological role of base methylation, a well-conserved modification of rRNA, is poorly understood. Here, wedemonstrate that a nucleolar factor, nucleomethylin (NML; also knownas RRP8), is required for the N1-methyladenosine (m1A) modificationin 28S rRNAs of human and mouse cells. NML also contributes to 60Sribosomal subunit formation. Intriguingly, NMLdepletion increases 60Sribosomal protein L11 (RPL11) levels in the ribosome-free fraction andprotein levels of p53 through an RPL11–MDM2 complex, whichactivates the p53 pathway. Consequently, the growth of NML-depletedcells is suppressed in a p53-dependent manner. These observationsreveal a new biological function of rRNA base methylation, which linksribosomal subunit formation to p53-dependent inhibition of cellproliferation in mammalian cells.

KEY WORDS: Nucleolar factor, NML, rRNA modification, m1A, p53,Cell proliferation

INTRODUCTIONRibosomal RNAs (rRNAs) function as scaffolds for ribosomalproteins and ribozymes for peptide bond formation (Nissen et al.,2000; Steitz and Moore, 2003). Biosynthesis of rRNA comprisestranscriptional and post-transcriptional modification of precursor(pre)-rRNAs as well as the processing of pre-rRNAs into mature28S, 18S and 5.8S rRNAs (Henras et al., 2015). Nucleotidemodifications of rRNA regulate the function and stability ofribosomes (Decatur and Fournier, 2002). There are three main typesof chemical modifications of rRNA – conversion of uridine topseudouridine (pseudouridylation); methylation of 2′-hydroxyls(2′-O-ribose methylation); and alteration of bases, most of whichundergo methylation at different positions (base methylation)(Decatur and Fournier, 2002). Of these modifications, basemethylations are the most conserved in terms of their totalnumber and position among living organisms, and are commonlyfound at approximately ten positions in eukaryotic rRNAs

(Hauenschild et al., 2015; Sharma and Lafontaine, 2015). Basemethylation is considered to expand the structural repertoire ofRNA, facilitating base stacking by increasing hydrophobicity andadjusting steric hindrance (Ishitani et al., 2008). To date, basemethylation of rRNA and the genes responsible for it (Bud23, Rrp8,Nop2, Rcm1, Bmt2, Bmt5 and Bmt6) have been reported inbudding yeast (Gigova et al., 2014; Peifer et al., 2013; Sharma et al.,2013; White et al., 2008). A very recent report has shown that yeastand worm homologs of human NSUN5 methylate C2278 in 25SrRNA of Saccharomyces cerevisiae and C2381 in 26S rRNA ofCaenorhabditis elegans; in addition, the reduction in the levels ofNSUN5 homologs decreases translational fidelity in yeast andincreases the lifespan of S. cerevisiae, C. elegans and Drosophilamelanogaster (Schosserer et al., 2015). The human ortholog of yeastBud23, WBSCR22 (Merm1), has also been shown to mediateN7-methylation of G1639 and 18S rRNA maturation in cells (Haaget al., 2015); however, little is known about the basemethyltransferase of rRNA in mammals.

In all organisms, ribosomes serve as the sole site of biologicprotein synthesis, and their biogenesis is strictly regulated. Themammalian ribosome comprises two subunits composed of rRNAsand ribosomal proteins (Anger et al., 2013; Khatter et al., 2015). Thesmall subunit (40S) is formed from a single molecule of rRNA (18SrRNA) and 33 ribosomal proteins (ribosomal proteins of the smallribosomal subunit; RPSs). Conversely, the large subunit (60S)contains three rRNA molecules (5S, 5.8S and 28S) and 47 ribosomalproteins (ribosomal proteins of the large subunit; RPLs).Interestingly, ribosome biogenesis defects are coupled to inhibitionof cell proliferation that is mediated by p53 (encoded by TP53)(Fumagalli et al., 2009; Sloan et al., 2013; Sulic et al., 2005). p53functions as a tumor suppressor, and its activation induces apoptosisand cell cycle arrest, resulting in cell growth suppression (Bieginget al., 2014; Brooks and Gu, 2010; Vousden, 2000). Together withthese reports, recent studies suggest that quantitative and qualitativechanges in ribosomes are associated with numerous physiologic andpathologic events (Barna et al., 2008; Belin et al., 2009, 2010;Figueiredo et al., 2015; Marcel et al., 2013). However, it remainsunclear whether and how the base methylation of rRNA regulatesribosomal functions and related biological events in mammals.

Nucleomethylin (NML; also known as RRP8) is a nucleolarprotein that binds to dimethyl lysine at position 9 of histone H3(H3K9me2) in cells and suppresses ribosomal DNA (rDNA)transcription in response to glucose deprivation (Grummt andLadurner, 2008; Murayama et al., 2008). NML contains theRossmann-fold methyltransferase-like domain in its C-terminalhalf, and binds to S-adenosyl-L-methionine (SAM), although NMLdoes not exhibit methyltransferase activity toward histones(Murayama et al., 2008). In contrast, the yeast NML homologRrp8 has been identified as a gene responsible for the SAM-dependent N1-methyladenosine (m1A) modification in the 25SReceived 19 November 2015; Accepted 29 April 2016

1Graduate School of Pharmaceutical Sciences, University of Tokyo, Bunkyo-ku,Tokyo 113-0033, Japan. 2Life Science Center, Tsukuba Advanced ResearchAlliance (TARA), University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan.3Graduate School of Life and Environmental Sciences, University of Tsukuba,Tsukuba, Ibaraki 305-8577, Japan. 4Laboratory for Genetic Code, Graduate Schoolof Life and Medical Sciences, Doshisha University, Kyotanabe, Kyoto 610-0394,Japan.*Present address: Laboratory for Genetic Code, Graduate School of Life andMedical Sciences, Doshisha University, Kyotanabe, Kyoto 610-0394, Japan.

‡Author for correspondence (nakajima@tara.tsukuba.ac.jp)

Y.N., 0000-0002-7357-2910

2382

© 2016. Published by The Company of Biologists Ltd | Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

rRNA (Peifer et al., 2013). However, the contribution of NML tom1A modification has yet to be elucidated. In this study,using mammalian cells, we reveal that NML is responsible form1A modification of 28S rRNA and contributes to the 60Sribosomal subunit formation, and that NML deficiency inhibits cellproliferation in a p53-dependent manner.

RESULTSNML is required for m1A modification in 28S rRNAIn yeast 25S rRNA, two sites with the m1A modification have beenidentified and, along with the neighboring sequences (Peifer et al.,2013), these sites are well conserved among eukaryotes, including inhumans andmice (Fig. 1A; Fig. S1; A645 andA2142 of 25S rRNA inSaccharomyces cerevisiae correspond to A1309 and A3625 of 28SrRNA inHomo sapiens, and A1136 and A3301 of 28S rRNA inMusmusculus, respectively). To confirm the interaction between NMLand 28S rRNA in human epithelial adenocarcinoma HeLa cells, weperformed RNA immunoprecipitation (RIP) experiments using anantibody against NML. In this assay, endogenousNML proteins wereprecipitated using an antibody against NML. RNAs were thenpurified from the precipitate, and 28S rRNA levels were quantified byperforming quantitative reverse-transcriptase (qRT)-PCR using aspecific primer set (Table S1). 28S rRNA was co-immunoprecipitated with NML from the whole cell (Fig. 1B) andnuclear extracts (Fig. 1C), indicating the association of NML with28S rRNA. By contrast, 5S rRNA was not significantly co-immunoprecipitated with NML (Fig. 1B,C). Next, to investigate theeffect of NML on rRNA methylation, we performed a method basedon site-specific semi-quantitative RT-PCR using HeLa cells in whichNML-knockdown (KD) constructs (shNML#1 and #2) were stablyexpressed (Fig. 1D, left panel). Methylation levels around A1309,which is methylated by Rrp8 in yeast, were significantly lower inshNML-transfected cells than in control cells expressing shRNAsagainst green fluorescent protein (shGFP) (Fig. 1D, right panel).However, methylation levels around A3625, which is methylated byBmt2 in yeast (Sharma et al., 2013), were very low in control cells[rRNA methylation levels around A1309 and A3625 (mean±s.d.)were 19.05±0.47 and 1.32±0.21, respectively. These values werecalculated so that the value obtainedwith the qRT-PCR reaction at thelow dNTP concentration was normalized to that obtained at the highdNTP concentration; see Materials and Methods.] and were notmarkedly changed by knockdown of NML (Fig. S2A). Consistently,methylation levels around A1136, which corresponds to A1309 of28S rRNA in humans, were reduced in NML-knockout (NML KO)immortalized mouse embryonic fibroblasts (MEFs), compared within NML wild-type (NML WT) MEFs (Fig. 1E; Fig. S2B). Primerextension analysis was also performed to confirm the modification atA1309 of HeLa cells and A1136 of immortalized MEFs in 28SrRNAs. Strong stop signals at positions A1309 and A1136 wereobserved in 28S rRNAs of shGFP and NMLWT cells (Fig. 1F). Bycontrast, these stop signals were reduced using 28S rRNAs of NMLKD and KO cells. Moreover, the reduction of methylation levelsaround A1136 was attenuated by the expression of FLAG- andhemagglutinin (HA) FLAG–HA-tagged NML, but not by theexpression of control EGFP (Fig. 1G; Fig. S2C). These resultsindicate that NML is involved in the modifications at A1309 andA1136 in 28S rRNA of human and mouse cells.To investigate whether NML regulates m1A modification of

28S rRNA, we then analyzed the base modifications of RNAsthat included a 28S rRNA fragment by performing liquid-chromatography-coupled mass spectrometry (LC-MS/MS). In NMLKO and KD cells, the peaks corresponding to m1Amodification were

reduced compared with those in NML WT and control KD cells(Fig. 2A,B). Moreover, RIP experiments using an antibody againstm1A revealed that m1A modifications around A1136 and A1309were significantly reduced by NML depletion (Fig. 2C,D). Theseresults indicate that NML is responsible for m1A modifications atpositions 1309 and 1136 in human and mouse 28S rRNA.

A previous study has revealed that the C-terminal region of NMLcontains a Rossman-fold methyltransferase-like domain, whichinteracts with the methyl donor SAM (Murayama et al., 2008)(Fig. 3A). To investigate whether this domain is indispensable forrRNA methylation, we performed a site-specific semi-quantitativeRT-PCR-based method using the immortalized MEFs that stablyexpressed FLAG–HA-tagged wild-type NML (NML-wt) ormutated NML constructs (NML-mt1 or NML-mt2) that lack theability to bind to SAM (Murayama et al., 2008) (Fig. 3A,B).Expression of NML-wt increased rRNA methylation levels aroundA1136 in NML KO cells (Fig. 3C). Conversely, no significantenhancement of rRNA methylation was induced by expression ofcontrol EGFP, NML-mt1 or NML-mt2 in NML KO cells. Theseresults suggest that NML regulates m1A modification in 28S rRNAthrough its methyltransferase-like domain.

NML contributes to the formation of the 60S ribosomalsubunitModifications of rRNA contribute to ribosomal subunit formationand association (Decatur and Fournier, 2002; Polikanovet al., 2015).To investigate the effect of knockdown of NML on the interactionbetween ribosomal subunits, we used bimolecular fluorescencecomplementation (BiFC) to visualize ribosomal subunit joining (Al-Jubran et al., 2013). The 40S ribosomal protein S18 (RPS18) and60S ribosomal protein L11 (RPL11), which are adjacent to eachother on the surface of each subunit, were tagged with the N- and C-terminal halves of the yellow fluorescent protein Venus, respectively(yielding S18-VN and L11-VC; Fig. 4A). Subunit interaction wasdetected as Venus signal (hereafter referred to as BiFC signal; seeMaterials andMethods).We validated this assay using fluorescence-activated cell sorting (FACS) and confirmed that the BiFC signalwas only detected when S18-VN and L11-VC were co-transfected(Fig. 4B). Moreover, the BiFC signal was dramatically reducedwhen cells were treated with puromycin, an aminoacyl-tRNA-likemolecule that causes 80S dissociation (Al-Jubran et al., 2013; Blobeland Sabatini, 1971) (Fig. 4B–D). These results suggest that the BiFCsignal derived from S18-VN and L11-VC expression is aconsequence of subunit interaction. Knockdown of NMLincreased the population of cells that emitted a low BiFC signaland decreased the proportion of cells that emitted a high BiFC signal(Fig. 4C). The median BiFC signals were also decreased byknockdown of NML (Fig. 4D). We then performed sucrose densitygradient centrifugation in the presence of EDTA, which dissociatesthe 80S complex into the 40S and 60S subunits (Peifer et al., 2013);the cell lysates were centrifuged and fractionated, and the ribosomalsubunit ratio was quantified by measuring the 18S and 28S rRNAlevels in each fraction (Fig. 5A). The ratio of 18S rRNA:28S rRNA(amount of 18S rRNA divided by the amount of 28S rRNA) inshNML cells was significantly higher than that of shGFP cells(Fig. 5B). Notably, in these cell lines, 18S and 28S rRNAs weredistributed to similar extents in fractions 4–8 and 9–15, respectively(Fig. 5A). Considering that this assay separates cellular organelles,including ribosomal subunits, depending on their individualdensities (mass/volume), these results indicate that knockdown ofNML results in fewer 60S subunits, rather than the formation ofabnormal subunits.

2383

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

Next, we tested whether NML depletion alters bulk proteintranslation by incorporating 35S-labeled methionine (35S-Met) intoproteins. Pretreatment with protein synthesis inhibitor cycloheximide(CHX) fully blocked amino acid incorporation into newly synthesized

proteins (Fig. 5C). InNMLKDcells, the rates of 35S-Met incorporationwere not markedly different from those in control cells. Overall, theseresults suggest that NML is involved in the formation of 60S ribosomalsubunits without affecting bulk protein synthesis.

Fig. 1. NML is required for modification in 28S rRNA. (A) Multiple sequence alignment of the regions neighboring two positions of N1-methyladenosine (m1A;A645 and A2142 in squares) of 25S rRNA (613–663 and 2128–2160) in Saccharomyces cerevisiae, of 28S rRNA (1295–1327 and 3611–3643) inHomo sapiensand of 28S rRNA (1122–1154 and 3287–3319) inMus musculus. Conserved nucleotides are marked with asterisks. (B,C) Interaction between rRNA and NML inwhole-cell extracts (B) and nuclear extracts (C) of HeLa cells. The binding of NML to 28S or 5S rRNAs was validated with RIP using an antibody against NML(αNML). An unconjugated affinity purified immunoglobulin (IgG) from mouse was used as a control. In the left panel in C, protein levels of NML in the indicatedfractions were determined using western blot. α-tubulin and histone H3 (H3) were also used as cytoplasmic and nuclear markers, respectively. (D) Methylationlevels at a region of 28S rRNA, including A1309, in HeLa cells expressing shRNAs targeting GFP (shGFP) or NML shNML#1 or shNML#2 were analyzed byperforming qRT-PCR (right panel). shGFP was used as a control shRNA. Protein levels of NML in the indicated cells were determined using western blot (leftpanel). (E) Methylation levels were analyzed by performing qRT-PCR (right panel) at a region of 28S rRNA, including A1136, in NML wild-type (WT) or knockout(KO) immortalized MEFs. Protein levels of NML in the indicated cells were determined using western blot (left panel). (F) Modification state at A1309 of human28S rRNA in shGFP- or shNML#2-expressing HeLa cells and at A1136 of mouse 28S rRNA in NMLWT or KO immortalized MEFs, analyzed by using primerextension tests. (G) Expression of FLAG and HA (FLAG–HA)-tagged NML recovered the methylation levels around A1136 of 28S rRNA in NMLWT or KOimmortalized MEFs, demonstrated using western blot (left panel) and qRT-PCR (right panel). All values are presented as mean±s.d. of three independentexperiments. **P<0.01; ***P<0.005; n.s., not significant (bootstrap and permutation tests) in B–E,G).

2384

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

Previous reports have shown that NML inhibits rDNAtranscription in human cells under glucose deprivation, but notunder conditions with 1 g/l glucose (Murayama et al., 2008; Yanget al., 2013) (Fig. S2D). Moreover, it has been reported that thedepletion of certain factors involved in ribosome biogenesis inhibitsrDNA transcription and/or pre-rRNA processing (Bousquet-Antonelli et al., 2000). We investigated the effect of knockdownof NML on rDNA transcription and processing under conditionswith 4.5 g/l glucose used in this study. In shNML-expressing HeLacells, mature and pre-rRNA levels were unchanged in comparisonwith those in control cells (Fig. S2D). Furthermore, analysis of pre-rRNA processing using metabolic labeling with 32P-orthophosphate

showed that there was no substantial change in the distributionpatterns on gels of pre-rRNA (47S and 45S, and 32S) and maturerRNA (28S and 18S) between shNML- and shGFP-expressingHeLa cells (Fig. S2E). These results indicate that NML affectsrRNA base methylation without altering rDNA transcription andpre-rRNA processing under normal glucose conditions.

NML depletion induces activation of the p53 pathwaythrough RPL11Location of ribosomal proteins outside of the ribosome increaseswhen ribosomal biogenesis is perturbed by nucleolar stress (Milianide Marval and Zhang, 2011; Zhang and Lu, 2009). Several

Fig. 2. NML is involved in m1A modification in 28S rRNA. (A,B) The amount of 28S rRNA m1A modified in NMLWT or KO immortalized MEFs (A) andshGFP or shNML#2 HeLa cells (B) was analyzed by performing LC-MS/MS. (C,D) m1A levels around A1136 in 28S rRNA of NMLWT or KO immortalized MEFs(C) and around A1309 in 28S rRNA of shGFP- or shNML-expressing HeLa cells (D) analyzed by RIP using an antibody against m1A (αm1A). An unconjugatedaffinity purified IgG from mouse was used as a control. Values are presented as mean±s.d. of three independent experiments. ***P<0.005; n.s., not significant(Student’s t-test) in C. **P<0.01; ***P<0.005; n.s., not significant (one-way ANOVA followed by Tukey’s test) in D.

Fig. 3. NML regulates 28S rRNAmethylation through its Rossman-foldmethyltransferase-like domain. (A) Schematic representation of wild-type (wt) NMLand point mutants in the Rossman-fold methyltransferase-like domain (mt1 and mt2). Conserved motifs in the SAM-methyltransferase are shown in red bars.(B) Protein levels of NML inNMLWTor KO immortalized MEFs expressing EGFP as a control. Expression of wild-type or mutated (mt1 or mt2) FLAG–HA-taggedNML was analyzed using western blot. (C) Methylation levels around A1136 of 28S rRNA in NML-KO immortalized MEFs, demonstrated by qRT-PCR analysis.Values are presented as mean±s.d. of three independent experiments. ***P<0.005; n.s., not significant (bootstrap and permutation tests).

2385

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

ribosomal proteins, including RPL11, bind to MDM2 and inhibitthe E3 ubiquitin ligase function of MDM2 toward p53, leading top53 stabilization and activation (Deisenroth and Zhang, 2010;Lohrum et al., 2003; Zhang et al., 2003). To investigate whetherNML is involved in p53 activation through ribosomal proteins(Fig. 6A), we analyzed the levels of RPL11 protein in the ribosome-free fraction by performing sucrose density gradient centrifugationin the absence of EDTA. The amount of ribosome-free RPL11 wasincreased in shNML cells (Fig. 6B; Fig. S3A), indicating elevatedRPL11 accumulation outside of the ribosome. Moreover, co-immunoprecipitation experiments showed that the interactionbetween RPL11 and MDM2 was increased in shNML cells(Fig. 6C). p53 activity in the immortalized MEFs and HeLa cellsis inhibited by SV40 large T-antigen and by E6 protein fromoncogenic HPV type 16, respectively (Hoppe-Seyler and Butz,1993; Kierstead and Tevethia, 1993). Thus, we used human coloncarcinoma HCT116 tp53 wild-type ( p53+/+) and null ( p53−/−)cells, and investigated the possibility that NML depletion inducesp53 activation through RPL11. In p53+/+ cells, knockdown of NMLincreased p53 protein levels, and this p53 accumulation wasabrogated by knockdown of RPL11 (Fig. 6D). The expression ofp53 target genes was induced by knockdown of NML in p53+/+

cells, but not in p53−/− cells, at both the mRNA and protein level

(Fig. 7). The activation of p53 target genes induced by NMLdepletion was also observed in normal MEFs, but not in theimmortalized MEFs (Fig. S3B). Taken together, these resultsindicate that NML knockdown increases the RPL11–MDM2interaction and induces p53 activation.

NML regulates cell proliferation in a p53-dependent mannerp53 contributes to the regulation of cell proliferation by inducingapoptosis and cell cycle arrest, or by inhibiting protein synthesis(Bieging et al., 2014; Brooks and Gu, 2010; Tilleray et al., 2006;Vousden, 2000). We investigated the effect of NML on these p53-dependent cellular events. A terminal deoxynucleotidyl transferasedUTP nick-end labeling (TUNEL) assay showed that the percentageof TUNEL-positive cells was significantly increased by NMLknockdown in p53+/+ cells, whereas this was not the case in p53−/−

cells (Fig. 8A; Fig. S4A). Moreover, cell cycle analysis revealed thatNML knockdown decreased the proportion of S-phase cells andincreased the proportion of G2- andM- (G2/M) phase cells in a p53-dependent manner 3 days after transfection with small interfering(si)RNAs (Fig. 8B; Fig. S4B). The rates of 35S-Met incorporationinto synthesized proteins were reduced by NML knockdown inp53+/+ cells (Fig. 8C). By contrast, in p53−/− cells, the rates ofincorporation of the labeled amino acid were not significantly

Fig. 4. The effect of knockdown of NML on the interaction between ribosomal subunits. (A) Crystal structure of the human ribosome (Protein DataBank ID,4V6X; Khatter et al., 2015). RPS18 and RPL11 and A1309 are highlighted in different colors. The BiFC model of the ribosome is also schematically represented.In this model, RPS18 and RPL11 are tagged with the N- and C-terminal parts of Venus, respectively (S18-VN and L11-VC). (B–D) The interaction betweenribosomal subunits in HeLa cells expressing shRNA targeting luciferase (shLuc) or NML (shNML#1 or #2) was analyzed by using BiFC following FACS analysis.shLuc was used as a control shRNA. S18-VN and/or L11-VC were co-expressed in cells with mCherry as a transfection control marker, and treated with or withoutpuromycin (puro). Venus and mCherry emissions (BiFC signals) for shLuc-expressing HeLa cells in the conditions indicated are plotted in B. The distribution andmedian of the BiFC signals are shown in C and D, respectively. Values are presented as mean±s.d. of three independent experiments. ***P<0.005 (one-wayanalysis of variance followed by Tukey’s test).

2386

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

changed by NML knockdown. Consistent with these results, cellproliferation was suppressed by NML knockdown in a p53-dependent manner (Fig. 8D). NML knockdown also significantlyreduced methylation levels around A1309, as well as interactionsbetween ribosomal subunits in both p53+/+ and p53−/− cells(Fig. 8E,F). Overall, our findings suggest that NML is involved inlarge ribosomal subunit formation throughm1Amodification of 28SrRNA, thereby regulating cell proliferation through the p53 pathway.

DISCUSSIONWe showed that the nucleolar protein NML regulates m1Amodification of 28S rRNA in a SAM-binding-domain-dependentmanner (Figs 1–3). Additionally, NML depletion decreases 60Ssubunit formation (Fig. 5A,B). Intriguingly, we observed that NMLdepletion enhances RPL11–MDM2 interactions and increases levelsof p53 protein, thereby activating the p53 pathway (Figs 6 and 7).Consequently, the proliferation of NML-depleted cells is suppressedby p53-dependent apoptosis, cell cycle arrest and protein synthesisinhibition (Fig. 8A–D). Therefore, our results suggest that NML isinvolved in them1Amodification of 28S rRNA and that a deficiencyin NML inhibits cell proliferation in a p53-dependent manner.NML was first identified as an H3K9me2-binding protein that

forms a complex with SIRT1 and SUV39H1 to suppress rDNAtranscription under glucose starvation (Murayama et al., 2008). Ourdata raise the question of how the molecular function ofNML switches from an epigenetic suppressor to a basemethyltransferase. Interestingly, it has been recently reported thatNML binds directly to rRNA under normal glucose conditions

(Yang et al., 2013). Conversely, glucose deprivation inhibits theinteraction between NML and rRNA, thereby enhancing therecruitment of SIRT1 to NML. Indeed, we confirmed that, at aglucose concentration of 4.5 g/l, NML interacts with 28S rRNA inHeLa cells (Fig. 1B,C) and that NML depletion does not change thelevel of rDNA transcription (Fig. S2D). These results suggest thatNML functions as a methyltransferase for rRNA under normalglucose conditions, whereas it functions as a suppressor of rDNAtranscription under glucose deprivation. Thus, NMLmightmodulatedistinct steps of ribosome biogenesis, involving rDNA transcriptionand rRNA methylation, in response to environmental changes.

Of the rRNA modifications, 2′-O-ribose methylation, whichaffects not only ribosome biogenesis but also ribosomal functions,including translational efficiency and fidelity (Decatur and Fournier,2002), is one of the best characterized. Furthermore, fibrillarin (FBL)has been identified as a 2′-O-ribose methyltransferase (Decatur andFournier, 2003; Lin et al., 2011). Recently, it has been reported thatp53 directly represses FBL transcription and decreases the 2′-O-ribose methylation of rRNAs, impairing the translational function ofribosomes (Marcel et al., 2013). Furthermore, FBL overexpressionfacilitates tumorigenesis and is associated with poor survival ofcancer-affected individuals, implying that rRNA modification playsa crucial role in cancer development. Indeed, levels of m1A-modifiednucleosides are elevated in the urine of individuals with cancer (Itohet al., 1992, 1988). Our results suggest that the NML-dependentm1Amodification of rRNA could influence cancer development throughthe p53 pathway. Therefore, potential roles of NML-dependentmodifications of rRNA in cancer development should be explored.

Fig. 5. NML is involved in theformation of 60S ribosomal subunit.(A,B) Ribosomal subunit formation inshGFP-, or shNML#1- or shNML#2-expressing HeLa cells were analyzed byperforming sucrose density gradientcentrifugation following qRT-PCR. 18Sand 28S rRNA levels in each fraction werenormalized against those in an input (A).The ratios between the total amounts of18S and 28S rRNAs are shown as subunitratios for the cells indicated (B). (C) Thetotal protein synthesis rate in the cellsindicated was measured based on35S-methionine incorporation (incorp.)into newly synthesized proteins.Cycloheximide (CHX) was used as apositive control. All values are presentedas mean±s.d. of three independentexperiments. *P<0.05; ***P<0.005; n.s.,not significant (one-way ANOVA followedby Tukey’s test) in B. ***P<0.005; n.s., notsignificant (bootstrap and permutationtests) in C.

2387

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

MATERIALS AND METHODSCell cultureHeLa cells were provided by the RIKEN Catalysis Research Center throughthe National Bio-Resource Project of the Ministry of Education, Culture,Sports, Science and Technology, Japan. Dr Akiko Murayama (University ofTsukuba, Tsukuba, Japan) gifted SV40 large-T-antigen-immortalized NMLWT and KOMEFs, which were generated from C57BL/6J background wild-type and NML-KO mice (Oie et al., 2014), respectively. HCT116 p53+/+ andp53−/− cells were kindly provided by Dr Bert Vogelstein (John HopkinsUniversity, Baltimore, MD) (Bunz et al., 1998). All animal experiments wereapproved and performed in accordancewith the guidelines for the care and useof laboratory animals at University of Tsukuba. ForMEF isolation, embryos at13.5 days post coitum from C57BL/6J mice (CLEA Japan, Tokyo, Japan)were used. After the removal of the head and visceral tissues, the remainingbodieswerewashed and dissociated.Cellswere plated ondishes and incubated37°C with 5% CO2. The next day, floating cells were removed by washing inPBS.MEFswere usedwithin three passages. Cells andMEFsweremaintainedin Dulbecco’s modified Eagle’s medium (DMEM; Gibco™, Thermo FisherScientific) supplemented with 10% fetal bovine serum (FBS; Gibco™,Thermo Fisher Scientific) and penicillin–streptomycin (Sigma-Aldrich).

AntibodiesAntibodies against the following proteins and epitopes were used in thisstudy: β-actin (Sigma-Aldrich; A5316; 1:5000); α-tubulin (Sigma-Aldrich;T5168; 1:5000); Histone H3 (Cell Signaling Technology; 9715; 1:3000);1-methyladenosine (m1A; MBL International, Nagoya, Japan; D345-3);RPL11 (Cell Signaling Technology; 14382; 1:2000); MDM2 (Santa CruzBiotechnology; sc-965; 1:500); p53 (DO-1; Santa Cruz Biotechnology; sc-126; 1:2000); 14-3-3σ (1.N.6; Abcam; ab14123; 1:2000); Bax (Abcam;ab7977; 1:1000); and an unconjugated affinity purified isotype controlimmunoglobulin (IgG) from mouse (Santa Cruz Biotechnology; sc-2025).

Previously reported anti-human and mouse NML antibodies were generated(1:1000) (Murayama et al., 2008; Oie et al., 2014).

RNA interferenceTo generate stable knockdown cell lines, cells were transfected with thepiGENE™ hU6 plasmid (iGENE Therapeutics, Tsukuba, Japan) containingsequences targeting NML, GFP or luciferase (Luc). The transfected cellswere selected with puromycin. The target sequences were: 5′-GCCGCTT-TGAGGATGTTCGAA-3′ for shNML#1; 5′-GGGTAGTACTACAAAT-GATCC-3′ for shNML#2 (Murayama et al., 2008); 5′-GGCTACGTCCA-GGAGCGCACC-3′ for shGFP; and 5′-GTGCGCTGCTGGTGCCAACC-C-3′ for shLuc.

For transient knockdown, cells were transfected with 20 nM of StealthRNAi™ short interfering RNA (siRNA) (Invitrogen) using Lipofectamine®

RNAiMAX (Invitrogen) according to the manufacturer’s protocol. Thetarget sequences were as follows: 5′-CCGCUUUGAGGAUGUUCGAA-CCUUU-3′ for siNML#1 (human); 5′-CCUCAUACAUUAAGCCGCA-AGCAGU-3′ for siNML#2 (human); 5′-CCAAACUCGGCUUUAAGA-UUAUCUA-3′ for siNML (mouse); 5′-ACACAUCGAUCUGGGUAUC-AAAUAU-3′ for siRPL11. Stealth RNAi™ Luciferase Reporter control(siLuc) was used as a negative control.

Plasmid construction and mutagenesisA plasmid containing mouse wild-type NML was constructed by insertingthe HindIII–NotI (blunt ended with Klenow)-digested fragment fromFLAG- and HA-tagged mouse NML in pcDNA3 (a gift from DrMurayama,University of Tsukuba, Tsukuba, Japan) into the XhoI (filled with Klenow)site of the plasmid pPB-CAG.EBNXN (kindly provided by Dr AllanBradley, Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, UK).Mutations were introduced by using site-directed mutagenesis and PCR.The primer sequences were as follows (altered sequences are underlined):

Fig. 6. NML depletion increases p53 protein levels through RPL11. (A) Schematic model of NML-dependent p53 accumulation through RPL11. In functionalNML cells, p53 is degraded byMDM2 (left pathway). Depletion of NML increases ribosome-free RPL11, which enhances the interaction of RPL11 with MDM2 andinduces p53 accumulation (right pathway). (B) Protein levels of RPL11 in the ribosome-free (free) or ribosomal (Ribo.) fraction of shGFP- or shNML#2-expressingHeLa cells were analyzed by performing sucrose gradient centrifugation followed by western blotting (top three rows). 18S and 28S rRNAs are shown asreferences for ribosome sedimentation (bottom row). (C) Interaction between RPL11 and MDM2 in shGFP- or shNML#2-expressing HeLa cells was analyzedusing immunoprecipitation (I.P.) with an antibody against MDM2 and western blotting. An unconjugated affinity purified IgG frommouse was used as a control forimmunoprecipitation. (D) Protein levels of p53 in HCT116 p53+/+ cells that had been treated with siRNA targeting luciferase (siLuc) orNML (siNML#1 or siNML#2)and/or RPL11 (siRPL11). Two days after siRNA transfection, the levels of the indicated proteins were analyzed using western blot. Representative western blotdata are shown on the left. Graphs are expressed as the fold change relative to the indicated protein levels in siLuc-transfected HCT116 p53+/+ cells. Values arepresented as mean±s.d. of three independent experiments. **P<0.01; ***P<0.005; n.s., not significant (bootstrap and permutation tests).

2388

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

mouse NML mt1 (G317D,G319R), 5′-GCTGACTTTGACTGTAGAGA-TTGCCGC-3′; mouse NML mt2 (G317Q), 5′-GCTGACTTTGGCTGTG-AAGATTGCCGC-3′.

Western blottingTo detect proteins using western blotting, cells were lysed with RIPAbuffer [20 mMTris-HCl (pH 7.5), 150 mMNaCl, 2 mMEDTA, 0.8%NonidetP-40 (NP-40), 0.1% sodium dodecyl sulfate (SDS) and 0.5% sodiumdeoxycholate]. Cell extracts were fractionated by performing SDS-PAGE andtransferred to a polyvinylidene difluoride membrane using transfer apparatusaccording to the manufacturer’s protocol (Bio-Rad Laboratories). Theantibodies used are described in the Antibodies section. Quantitative analysesof western blot data were performed using Multi Gauge version 3.0 software(Fujifilm). Each protein level was normalized to the protein levels of β-actin.

Co-immunoprecipitationCells were lysed in RIPA buffer supplemented with EDTA-free proteaseinhibitor cocktail (Nacalai Tesque, Kyoto, Japan). Equal amounts ofextracted proteins were resuspended in TNE buffer [10 mM Tris-HCl (pH7.5), 150 mM NaCl, 1 mM EDTA, 1% NP-40] supplemented with EDTA-free protease inhibitor cocktail, and were immunoprecipitated with theindicated antibodies (1.8 μg/ml) and Dynabeads conjugated to protein G(Invitrogen). Bound proteins were then analyzed by using western blot.

qRT-PCRTo quantify rRNA methylation levels, we performed a site-specific rRNAmethylation assay based on quantitative reverse-transcription (qRT)-PCR

(Belin et al., 2009; Marcel et al., 2013). Total RNAwas extracted and purifiedusing Sepasol-RNA I Super G (Nacalai Tesque) according to themanufacturer’s instructions. Aliquots of total RNA (500 ng) were reversetranscribed using ReverTra Ace® (Toyobo, Osaka, Japan) and 1 µM of eachreverse primer targeting a sequence downstream to a specific methylation site,with either a low (2.5 µM) or high (250 µM) deoxy nucleoside triphosphate(dNTP) concentration, according to the manufacturer’s instructions. qRT-PCR was performed with SYBR® Premix Ex Taq™ II (Takara Bio, Otsu,Japan) using a Thermal Cycler Dice™ Real-Time System (Takara Bio). Theamount of methylation was calculated following the function 2(CTlow−CThigh),where the CT (threshold cycle) value obtained with the qRT-PCR reaction atthe low dNTP concentrationwas normalized to that obtained at the high dNTPconcentration. We used the following primer sets: hA1309_RT andhA1309_Fw/Rv; hA3625_RT and hA3625_Fw/Rv; mA1136_RT andmA1136_Fw/Rv; and mA3301_RT and mA3301_Fw/Rv (Table S1).

To quantify rRNA or mRNA levels, aliquots of total RNA (500 ng) werereverse transcribed with a pd(N)6 random primer and 250 µM of dNTPs.qRT-PCR was performed with primers for the indicated rRNA or genes(Table S1). The expression level of each gene in human and mouse cells wasnormalized to the mRNA levels of genes encoding human β-actin (ACTB)and mouse cyclophilin (Ppia), respectively.

RNA immunoprecipitationRIP was performed as previously described (Lin et al., 2011) with minormodifications. For RIP against NML using nuclear lysates, cells wereharvested and resuspended in 450 μl buffer A [10 mM HEPES (pH 7.5),10 mM KCl, 0.1 mM EDTA] (Dignam et al., 1983) and kept on ice for

Fig. 7. NML depletion activates the p53 pathway. (A) mRNA levels of p53 target genes in HCT116 p53+/+ or p53−/− cells that had been treated with siLuc orsiNML#1 or siNML#2. Three days after siRNA transfection, the levels of the indicated mRNAwere analyzed by performing qRT-PCR. Graphs are expressed asthe fold change relative to corresponding mRNA levels in siLuc-transfected HCT116 p53+/+ or p53−/− cells. (B) Protein levels of p53 targets in HCT116 p53+/+ orp53−/− cells that had been treated with siLuc or siNML#1. Three days after siRNA transfection, the indicated protein levels were analyzed using western blot.Representative western blot data are shown in the upper panel. Graphs show the fold change relative to the indicated protein levels in siLuc-transfected HCT116p53+/+ cells (lower panel). All values are presented as mean±s.d. of three independent experiments. *P<0.05; **P<0.01; ***P<0.005; n.s., not significant(bootstrap and permutation tests).

2389

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

15 min. Fifty microliters of 5% NP-40 in buffer A were added. Nuclei werepelleted by centrifugation at 800 g for 5 min. Then, whole cells or nucleiwere suspended in 200 μl cold lysis buffer [10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (pH 7.5), 300 mM KCl, 5 mM MgCl2, 0.5%NP-40, 1 mM dithiothreitol, EDTA-free protease inhibitor cocktail (Nacalai

Tesque) and 100 U/ml RNase Inhibitor (Toyobo]. Whole-cell or nuclearextracts were centrifuged at 15,000 g for 15 min at 4°C, and the supernatantwas pre-cleared using protein-G–Sepharose® beads (GE Healthcare) andtreated with DNase I (Wako Pure Chemical Industries, Osaka, Japan) (60 U/100 μl lysate) for 30 min on ice. The samples (2‒10 μg) were incubated with

Fig. 8. NML regulates cell proliferation in a p53-dependent manner. (A) Apoptosis assay of HCT116 p53+/+ or p53−/− cells that had been treated with siLuc, orsiNML#1 or siNML#2. Three days after siRNA transfection, the cells indicated were analyzed using TUNEL staining. For each group, cells were counted in threerandomly selected fields (more than 100 cells/field) per experiment. (B) Cell cycle analysis of HCT116 p53+/+ or p53−/− cells that had been treated with siLuc, orsiNML#1 or siNML#2. Three days after siRNA transfection, the indicated cells were analyzed using EdU staining following FACS analysis (left panel). Thepercentages of the cell population in the S-phase were compared between the indicated cells (right panel). (C) The total protein synthesis rates in the cellsindicated were measured based on 35S-methionine incorporation (incorp.) into newly synthesized proteins 3 days after siRNA transfection. (D) Phase-contrastimages and growth curve of HCT116 p53+/+ or p53−/− cells that had been treated with siLuc, or siNML#1 or siNML#2. The phase-contrast images (left panel) weretaken three days after siRNA transfection. Scale bar: 100 µm. Living cells were counted at the times indicated after siRNA transfection using Trypan Blue staining(right panel). The statistical significance between siLuc- and siNML#1- or siNML#2-expressing cells is shown by * and †, respectively. (E) Methylation levels at theregion of 28S rRNA that includes A1309 in HCT116 p53+/+ or p53−/− cells that had been treated with the indicated siRNAs. Three days after siRNA transfection,methylation levels were analyzed by performing qRT-PCR. (F) Themedian BiFC signals in HCT116 p53+/+ or p53−/− cells that had been treated with the indicatedsiRNAs. Three days after siRNA transfection, BiFC signals were analyzed by performing FACS. All values are presented as mean±s.d. of three independentexperiments in A–C,E,F and of three wells in D from a representative experiment that was performed at least three times with similar results. * or †P<0.05;**P<0.01; *** or †††P<0.005; n.s., not significant (one-way ANOVA followed by Tukey’s test) in A,B,D,F. **P<0.01; ***P<0.005; n.s., not significant (bootstrap andpermutation tests) in C,E.

2390

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

pre-binding antibodies (5 μg) and protein-G–Sepharose® beads in 1000 μlNT2 buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM MgCl2 and0.05% NP-40] including 100 U of RNase inhibitor, 10 mM dithiothreitoland 20 mM EDTA overnight at 4°C. The beads were washed twice withNT2 buffer and resuspended in 100 μl of NT2 buffer. The suspension wasmixed with 100 μl of proteinase K buffer [30 μg proteinase K, 20 mM Tris-HCl (pH 7.8), 10 mM EDTA and 1.0% SDS] and incubated for 30 min at55°C. After centrifugation, RNAs were purified from the supernatant usingphenol–chloroform–isoamyl-alcohol and analyzed by performing qRT-PCR. Whole-cell or nuclear extracts were used as input samples fornormalization of qRT-PCR data.

For RIP against m1A, total RNAs were extracted using Sepazol-RNA ISuper G and treated with DNase I. Purified RNAs were fragmented using anNEBNext® Magnesium RNA Fragmentation Module (New EnglandBiolabs). Fragmented RNAs were precipitated with ethanol after adding20 μg of glycogen and resuspended in 100 µl of IPP buffer [10 mM Tris-HCl (pH 7.5), 150 mM NaCl and 0.1% NP-40]. RNA samples were addedto 100 µl of an antibody mixture that included IPP buffer and 50 μl ofprotein-G–Sepharose® beads. The mixture was rotated overnight at 4°C.The beads were washed twice with IPP buffer. Immunoprecipitated RNAswere purified by phenol–chloroform–isoamyl-alcohol extraction, andanalyzed using qRT-PCR. Total RNAs were used as input samples fornormalization of qRT-PCR data.

Primer extensionPrimer extension was performed as per previously described protocols(Cozen et al., 2015; Sharma et al., 2013) with minor modifications. Onemicromole of DNA primer was 32P-5′-terminally labeled by incubation in afinal volume of 25 μl with 15 μCi γ-[32P]ATP and 9 U of polynucleotidekinase in protruding kinase buffer (Toyobo). The reaction was incubated at37°C for 1 h. The reaction mixture was then purified using Illustra™MicroSpin™ G-25 columns (GE Healthcare). For the extension reaction,0.5 μl of the 32P-5′-phosphorylated primer was annealed by heating for3 min at 95°C, followed by cooling on ice using 0.2 μg total RNA in 4.75 μlRT buffer (Toyobo) containing 2.5 μM dNTPs. Annealed primers wereextended using 0.25 μl ReverTra Ace® (25 U) for 1 h at 42°C, stopped bythe addition of 5 μl formamide loading dye and frozen at −80°C. Primerextension products were resolved by performing electrophoresis on 8%polyacrylamide gels containing 4 M urea. Gels were dried and exposed toAmersham Hyperfilm™ ECL (GE Healthcare). DNA primer sequences(hA1309_RT and mA1136_RT) used for primer extension are listed inTable S1.

Quantification of methylated ribonucleosides with LC-MS/MS28S rRNA was isolated from total RNA by performing agarose gelelectrophoresis and using the NucleoSpin® Gel and PCR Clean-up system(Macherey-Nagel, Düren, Germany) according to the instruction manual.After denaturation by heating at 100°C for 3 min, 1–2 µg of 28S rRNAwasimmediately chilled on ice. Denatured RNA was hydrolyzed by 1 U ofnuclease P1 (Wako Pure Chemical Industries) in 10 mM ammonium acetatebuffer (pH 5.3) at 45°C for 2 h, subsequently dephosphorylated byincubating with phosphodiesterase I (0.0002 U, Sigma-Aldrich), alkalinephosphatase (0.3 U, Toyobo) and 0.1 volume of 1 M ammoniumbicarbonate buffer (pH 7.9) at 37°C for 2 h, as described previously(Crain, 1990). Ten pmol of br5U (Tokyo Chemical Industry, Tokyo, Japan)were added into the digestion mixture as an internal standard; the enzymeswere subsequently removed by acetone precipitation. The supernatant wasleft to evaporate, and the ribonucleoside powder was dissolved with 15 µl ofHPLC-grade water (Wako Pure Chemical Industries).

Individual ribonucleosides were quantified using a Shimadzu Nexera™UHPLC system coupled to LCMS-8050™ triple quadrupole massspectrometer (Shimadzu, Kyoto, Japan). Reverse-phase HPLC separationwas carried out using a Kinetex™ 2.6 µm C18 column (2.1×150 mm,Phenomenex, Torrance, CA, USA) with a VANGUARD™ Pre-Column(2.1×5 mm, Waters Co., Milford, MA, USA) at 30°C with 80% acetonitrilein 0.1% formic acid (solvent B) gradient in 0.1% formic acid (solvent A) at aflow rate of 0.2 ml/min. The abovementioned water-dissolved sample (5 µl)was injected and eluted with the following gradient elution: initial isocratic

elution with 0% solvent B for 3 min, followed by linear gradient elutionfrom 0 to 8% B until 19 min, jumping to 100% B within 2 min and holdingthe status until 25.5 min. The column was then subsequently returned to theinitial conditions within 1 min and equilibrated for 4.5 min before the nextsample injection. For determining elution positions of ribonucleosides onthe chromatogram, standard chemicals of m1A (Santa Cruz Biotechnology),N6-methyladenosine (m6A) (Carbosynth, Compton, Berkshire, UK) andbr5U were used, and their retention times were revealed as 2.23 min,15.1 min and 11.58 min, respectively.

The mass spectrometer was operated with an ion-spray source at 300°C inpositive ion mode, with unit resolution for Q1 and Q3, and other optimizedparameters: interface voltage, 4.0 kV; interface current, 0.1 µA; flow rate ofnebulizer gas, 3 l/min; flow rate of heating gas, 10 l/min; flow rate of dryinggas, 10 l/min; collision gas (Ar), 270 kPa; desolvation line temperature,250°C; heat block temperature, 400°C; conversion dynode potential, 10 kV;detector potential, 2.44 kV. Multiple reaction monitoring was used fordetection of nucleosides with a dwell time of up to 100 ms. Q1 was set totransmit the parental ions MH+ atm/z 282.1, 282.1 and 324.6 for m1A, m6Aand br5U, respectively. The daughter ions were monitored in Q3 at m/z150.1, 150.1 and 193.05 for m1A, m6A and br5U, respectively. Linearcalibration curves were obtained daily.

All of the solvents and reagents used in this analysis were HPLC grade.Instrument control and data processing were performed using theLabSolutions LCMS (Ver.5.60) software (Shimadzu Co.).

Bimolecular fluorescence complementationThe BiFC assay was performed as described previously (Al-Jubran et al.,2013). To generate plasmids expressing BiFC-tagged ribosomal proteins,the sequences of RPS18 and RPL11were PCR-amplified from the cDNA ofHeLa cells using primers tagged with SacI and XbaI or ApaI and KpnI,respectively. RPS18 and RPL11 segments were cloned into pBiFC-VN173(Addgene, 22010) and pBiFC-VC155 (Addgene, 22011), respectively(Shyu et al., 2008). Using Lipofectamine® LTX (Invitrogen), these BiFCplasmids were co-transfected with a pmCherry-N1 plasmid (ClontechLaboratories), which was used as an internal control. Twenty-four hoursafter transfection, cells were further cultured with or without 100 µg/mlpuromycin for 24 h. The cells were washed twice with STM buffer[phosphate-buffered saline (PBS) with 2% FBS] and resuspended in 500 μlof STM buffer. The fluorescence intensities of Venus and mCherry weremeasured by flow cytometry using a BD FACSAria™ II (BD Biosciences).The BiFC signal emitted by a cell was calculated using the followingfunction: (fluorescent intensity of Venus)/(fluorescent intensity ofmCherry), as previously described (Hu et al., 2002).

rRNA processingCells were cultured with methionine-free medium for 1 h, and then withmethionine-free medium containing 20 µCi/ml 32P-labeled inorganicphosphate for 1 h (Sloan et al., 2013). The metabolic labeling media wereremoved, and the cells were cultured with normal medium for 3 h. RNAwasextracted using a FastPure™RNAKit (Takara Bio) and analyzed by glyoxalagarose gel electrophoresis. Total RNA was visualized using ethidiumbromide staining. The intensity of each band was analyzed with the MultiGauge version 3.0 (Fujifilm).

Sucrose density gradient centrifugationSucrose density gradient centrifugation was performed as describedpreviously (Peifer et al., 2013; Tuorto et al., 2012).

To investigate ribosomal subunit formation, cells were treated with100 µg/ml cycloheximide (CHX) in culture medium for 5 min at 37°C. Thecells were washed twice with PBS containing CHX (100 µg/ml) and lysedwith lysis buffer [15 mMTris-HCl (pH 7.5), 300 mMNaCl, 25 mMEDTA,1% Triton X-100, 0.5 mg/ml heparin, 100 µg/ml CHX, 1× EDTA-freeprotease inhibitor cocktail (Nacalai Tesque) and 100 U/ml RNase inhibitor(Toyobo)]. The lysate was centrifuged at 9300 g for 10 min at 4°C andloaded onto a linear 20–50% sucrose gradient buffer in 15 mM Tris-HCl(pH 8.0), 25 mM EDTA and 300 mM NaCl. Centrifugation was conductedat 40,000 rpm for 2 h at 4°C in a SW-41 Ti rotor (Beckman Coulter), and

2391

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

fractions were collected from the top of the gradient (#1 to #19). The 18Sand 28S rRNA levels in the fractions #1 to #18 were quantified byperforming qRT-PCR and normalized against those rRNA levels in an inputcontrol. The ratio of 18S rRNA:28S rRNA (18S rRNA divided by 28SrRNA) was calculated as the small:large ribosomal subunit ratio.

To obtain the ribosomal or ribosome-free fractions, we performed sucrosedensity gradient centrifugation using lysis and gradient buffers containing15 mM MgCl2 instead of 25 mM EDTA. 18S and 28S rRNAs, and β-actinproteins were used as indicators for the ribosomal and ribosome-freefractions, respectively (Lee et al., 2013; Morello et al., 2011).

Analysis of protein synthesis using 35S-methionine incorporationThis experiment was performed as previously described (Itani et al.,2003; Mieulet et al., 2007). Cells were plated onto 12-well plates incomplete medium and transfected with or without siRNAs. One day aftercell seeding or 3 days after transfection, the culture medium was switchedto methionine-free DMEM with or without 50 μg/ml CHX for 30 minat 37°C, and cells were incubated with the same medium containing20 μCi/ml 35S-methionine (PerkinElmer) for 2 h at 37°C. Cells were thenwashed twice with ice-cold PBS and solubilized in 50 µl of RIPA buffercontaining protease inhibitors. Aliquots of the supernatant were analyzedusing a liquid scintillation counter (Beckman Coulter) and normalizedagainst the total protein content.

5-ethynyl-2′-deoxyuridine staining for cell cycle analysisFor 5-ethynyl-2′-deoxyuridine (EdU) staining, cells were plated (1×105

cells/well in a 6-well plate) and cultured for the indicated number of daysafter siRNA transfection. Then, this experiment was conducted using aClick-iT® EdU Imaging Kit (Invitrogen), according to the manufacturer’sprotocol. First, cells were labeled with 10 µM EdU (Invitrogen) in culturemedium for 1 h at 37°C. For subsequent DNA staining, EdU-stained cellswere incubated with 2 µg/ml Hoechst 33342 (Sigma-Aldrich) for 10 min.Then the cells were washed with PBS and subjected to FACS by using flowcytometry in a BD FACSAria™ II instrument (BD Biosciences).

Terminal deoxynucleotidyl transferase dUTP nick-end labelingassayCells were plated onto poly-D-lysine-coated (BD Biosciences) 8-wellchamber slides (1.6×104–2.4×104 cells/well) and cultured for 1, 2, 3 or4 days after siRNA transfection. Then, the cells were fixed with 4%paraformaldehyde in PBS for 25 min at 4°C, and in situ detection ofapoptotic cells was performed using DeadEnd™ Fluorometric TUNELSystem (Promega), according to the manufacturer’s instructions. Nucleiwere then counterstained using 2 µg/ml Hoechst 33342. The cells wereexamined using the fluorescent microscope imaging system Biorevo BZ-9000 (Keyence, Osaka, Japan), and TUNEL-positive cells were quantifiedin three fields per sample. The percentage of TUNEL-positive cells wascalculated as follows: (TUNEL-positive/total cells)×100.

Cell proliferation assayWe seeded 1×105 cells/well in 6-well plates and transfected them with theindicated siRNA. The cells were then trypsinized into single-cellsuspensions and automatically counted using a TC10™ Automated CellCounter (Bio-Rad Laboratories) on days 1, 2, 3 and 4, where the first dayafter seeding was day 0.

StatisticsThe unpaired Student’s t-test was used to compare two groups, and one-wayANOVA followed by Tukey’s post-hoc test was used to compare multiplegroups. The statistical analysis of fold-change data was performed bybootstrap and permutation tests using the web application BootstRatio(Cleries et al., 2012).

AcknowledgementsWe are thankful to Dr Shin-ichi Kashiwabara (University of Tsukuba, Tsukuba,Japan) for his technical advice of sucrose density gradient centrifugation and primerextension assay. We are also grateful to the members of Shimizu and Fukamizulaboratories for technical advice and useful discussion.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsT.W., Y.N., W.Y. and N.N. designed and performed the experiment. T.W. and Y.N.wrote the manuscript. K.K., A.K., T.S. and A.F. supervised the project and helpedprepare the manuscript.

FundingThis work was supported byGrant-in-Aid for Scientific Research on Innovative Areas(Japan Society for the Promotion of Science) [grant numbers 26116725 to A.K.,23116007 to T.S. and 23116004 to A.F.]; and the Open Innovation Core (OIC)Project (Y.N.; a member of OIC) of Life Science Center, Tsukuba AdvancedResearch Alliance, University of Tsukuba, Japan.

Supplementary informationSupplementary information available online athttp://jcs.biologists.org/lookup/doi/10.1242/jcs.183723.supplemental

ReferencesAl-Jubran, K., Wen, J., Abdullahi, A., Roy Chaudhury, S., Li, M., Ramanathan,

P., Matina, A., De, S., Piechocki, K., Rugjee, K. N. et al. (2013). Visualization ofthe joining of ribosomal subunits reveals the presence of 80S ribosomes in thenucleus. RNA 19, 1669-1683.

Anger, A. M., Armache, J.-P., Berninghausen, O., Habeck, M., Subklewe, M.,Wilson, D. N. andBeckmann, R. (2013). Structures of the human andDrosophila80S ribosome. Nature 497, 80-85.

Barna, M., Pusic, A., Zollo, O., Costa, M., Kondrashov, N., Rego, E., Rao, P. H.and Ruggero, D. (2008). Suppression of Myc oncogenic activity by ribosomalprotein haploinsufficiency. Nature 456, 971-975.

Belin, S., Beghin, A., Solano-Gonzalez, E., Bezin, L., Brunet-Manquat, S.,Textoris, J., Prats, A.-C., Mertani, H. C., Dumontet, C. and Diaz, J.-J. (2009).Dysregulation of ribosome biogenesis and translational capacity is associatedwith tumor progression of human breast cancer cells. PLoS ONE 4, e7147.

Belin, S., Kindbeiter, K., Hacot, S., Albaret, M. A., Roca-Martinez, J.-X.,Therizols, G., Grosso, O. and Diaz, J.-J. (2010). Uncoupling ribosomebiogenesis regulation from RNA polymerase I activity during herpes simplexvirus type 1 infection. RNA 16, 131-140.

Bieging, K. T., Mello, S. S. and Attardi, L. D. (2014). Unravelling mechanisms ofp53-mediated tumour suppression. Nat. Rev. Cancer 14, 359-370.

Blobel, G. and Sabatini, D. (1971). Dissociation of mammalian polyribosomes intosubunits by puromycin. Proc. Natl. Acad. Sci. USA 68, 390-394.

Bousquet-Antonelli, C., Vanrobays, E., Gelugne, J.-P., Caizergues-Ferrer, M.and Henry, Y. (2000). Rrp8p is a yeast nucleolar protein functionally linked toGar1p and involved in pre-rRNA cleavage at site A2. RNA 6, 826-843.

Brooks, C. L. and Gu, W. (2010). New insights into p53 activation. Cell Res. 20,614-621.

Bunz, F., Dutriaux, A., Lengauer, C., Waldman, T., Zhou, S., Brown, J. P.,Sedivy, J. M., Kinzler, K.W. and Vogelstein, B. (1998). Requirement for p53 andp21 to sustain G2 arrest after DNA damage. Science 282, 1497-1501.

Cleries, R., Galvez, J., Espino, M., Ribes, J., Nunes, V. and de Heredia, M. L.(2012). BootstRatio: a web-based statistical analysis of fold-change in qPCR andRT-qPCR data using resampling methods. Comput. Biol. Med. 42, 438-445.

Cozen, A. E., Quartley, E., Holmes, A. D., Hrabeta-Robinson, E., Phizicky, E. M.and Lowe, T. M. (2015). ARM-seq: AlkB-facilitated RNA methylation sequencingreveals a complex landscape of modified tRNA fragments. Nat. Methods 12,879-884.

Crain, P. F. (1990). Preparation and enzymatic hydrolysis of DNA and RNA for massspectrometry. Methods Enzymol. 193, 782-790.

Decatur, W. A. and Fournier, M. J. (2002). rRNA modifications and ribosomefunction. Trends Biochem. Sci. 27, 344-351.

Decatur, W. A. and Fournier, M. J. (2003). RNA-guided nucleotide modification ofribosomal and other RNAs. J. Biol. Chem. 278, 695-698.

Deisenroth, C. and Zhang, Y. (2010). Ribosome biogenesis surveillance: probingthe ribosomal protein-Mdm2-p53 pathway. Oncogene 29, 4253-4260.

Dignam, J. D., Lebovitz, R. M. and Roeder, R. G. (1983). Accurate transcriptioninitiation by RNA polymerase II in a soluble extract from isolated mammaliannuclei. Nucleic Acids Res. 11, 1475-1489.

Figueiredo, V. C., Caldow, M. K., Massie, V., Markworth, J. F., Cameron-Smith,D. and Blazevich, A. J. (2015). Ribosome biogenesis adaptation in resistancetraining-induced human skeletal muscle hypertrophy. Am. J. Physiol. Endocrinol.Metab. 309, E72-E83.

Fumagalli, S., Di Cara, A., Neb-Gulati, A., Natt, F., Schwemberger, S., Hall, J.,Babcock, G. F., Bernardi, R., Pandolfi, P. P. and Thomas, G. (2009). Absenceof nucleolar disruption after impairment of 40S ribosome biogenesis reveals anrpL11-translation-dependent mechanism of p53 induction. Nat. Cell Biol. 11,501-508.

2392

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience

Gigova, A., Duggimpudi, S., Pollex, T., Schaefer, M. andKos,M. (2014). A clusterof methylations in the domain IV of 25S rRNA is required for ribosome stability.RNA 20, 1632-1644.

Grummt, I. and Ladurner, A. G. (2008). A metabolic throttle regulates theepigenetic state of rDNA. Cell 133, 577-580.

Haag, S., Kretschmer, J. and Bohnsack, M. T. (2015). WBSCR22/Merm1 isrequired for late nuclear pre-ribosomal RNA processing and mediates N7-methylation of G1639 in human 18S rRNA. RNA 21, 180-187.

Hauenschild, R., Tserovski, L., Schmid, K., Thuring, K., Winz, M.-L., Sharma,S., Entian, K.-D., Wacheul, L., Lafontaine, D. L. J., Anderson, J. et al. (2015).The reverse transcription signature of N-1-methyladenosine in RNA-Seq issequence dependent. Nucleic Acids Res. 43, 9950-9964.

Henras, A. K., Plisson-Chastang, C., O’Donohue, M.-F., Chakraborty, A. andGleizes, P.-E. (2015). An overview of pre-ribosomal RNA processing ineukaryotes. Wiley Interdiscip. Rev. RNA 6, 225-242.

Hoppe-Seyler, F. and Butz, K. (1993). Repression of endogenous p53transactivation function in HeLa cervical carcinoma cells by humanpapillomavirus type 16 E6, human mdm-2, and mutant p53. J. Virol. 67,3111-3117.

Hu, C.-D., Chinenov, Y. and Kerppola, T. K. (2002). Visualization of interactionsamong bZIP and Rel family proteins in living cells using bimolecular fluorescencecomplementation. Mol. Cell 9, 789-798.

Ishitani, R., Yokoyama, S. and Nureki, O. (2008). Structure, dynamics, andfunction of RNA modification enzymes. Curr. Opin. Struct. Biol. 18, 330-339.

Itani, O. A., Cornish, K. L., Liu, K. Z. and Thomas, C. P. (2003). Cycloheximideincreases glucocorticoid-stimulated alpha -ENaCmRNA in collecting duct cells byp38 MAPK-dependent pathway. Am. J. Physiol. Renal Physiol. 284, F778-F787.

Itoh, K., Mizugaki, M. and Ishida, N. (1988). Preparation of a monoclonal antibodyspecific for 1-methyladenosine and its application for the detection of elevatedlevels of 1-methyladenosine in urines from cancer patients. Jpn. J. Cancer Res.79, 1130-1138.

Itoh, K., Ishiwata, S., Ishida, N. and Mizugaki, M. (1992). Diagnostic use of anti-modified nucleoside monoclonal antibody. Tohoku J. Exp. Med. 168, 329-331.

Khatter, H., Myasnikov, A. G., Natchiar, S. K. and Klaholz, B. P. (2015). Structureof the human 80S ribosome. Nature 520, 640-645.

Kierstead, T. D. and Tevethia, M. J. (1993). Association of p53 binding andimmortalization of primary C57BL/6 mouse embryo fibroblasts by using simianvirus 40 T-antigen mutants bearing internal overlapping deletion mutations.J. Virol. 67, 1817-1829.

Lee, M. S., Kim, B., Oh, G. T. and Kim, Y.-J. (2013). OASL1 inhibits translation ofthe type I interferon-regulating transcription factor IRF7. Nat. Immunol. 14,346-355.

Lin, J., Lai, S., Jia, R., Xu, A., Zhang, L., Lu, J. and Ye, K. (2011). Structural basisfor site-specific ribose methylation by box C/D RNA protein complexes. Nature469, 559-563.

Lohrum, M. A. E., Ludwig, R. L., Kubbutat, M. H. G., Hanlon, M. and Vousden,K. H. (2003). Regulation of HDM2 activity by the ribosomal protein L11. CancerCell 3, 577-587.

Marcel, V., Ghayad, S. E., Belin, S., Therizols, G., Morel, A.-P., Solano-Gonzalez, E., Vendrell, J. A., Hacot, S., Mertani, H. C., Albaret, M. A. et al.(2013). p53 acts as a safeguard of translational control by regulating fibrillarin andrRNA methylation in cancer. Cancer Cell 24, 318-330.

Mieulet, V., Roceri, M., Espeillac, C., Sotiropoulos, A., Ohanna, M., Oorschot,V., Klumperman, J., Sandri, M. and Pende, M. (2007). S6 kinase inactivationimpairs growth and translational target phosphorylation in muscle cellsmaintaining proper regulation of protein turnover. Am. J. Physiol. Cell Physiol.293, C712-C722.

Miliani de Marval, P. L. and Zhang, Y. (2011). The RP-Mdm2-p53 pathway andtumorigenesis. Oncotarget 2, 234-238.

Morello, L. G., Hesling, C., Coltri, P. P., Castilho, B. A., Rimokh, R. and Zanchin,N. I. T. (2011). The NIP7 protein is required for accurate pre-rRNA processing inhuman cells. Nucleic Acids Res. 39, 648-665.

Murayama, A., Ohmori, K., Fujimura, A., Minami, H., Yasuzawa-Tanaka, K.,Kuroda, T., Oie, S., Daitoku, H., Okuwaki, M., Nagata, K. et al. (2008).Epigenetic control of rDNA loci in response to intracellular energy status.Cell 133,627-639.

Nissen, P., Hansen, J., Ban, N., Moore, P. B. and Steitz, T. A. (2000). Thestructural basis of ribosome activity in peptide bond synthesis. Science 289,920-930.

Oie, S., Matsuzaki, K., Yokoyama, W., Tokunaga, S., Waku, T., Han, S.-I.,Iwasaki, N., Mikogai, A., Yasuzawa-Tanaka, K., Kishimoto, H. et al. (2014).Hepatic rRNA transcription regulates high-fat-diet-induced obesity. Cell Rep. 7,807-820.

Peifer, C., Sharma, S., Watzinger, P., Lamberth, S., Kotter, P. and Entian, K.-D.(2013). Yeast Rrp8p, a novel methyltransferase responsible for m1A 645 basemodification of 25S rRNA. Nucleic Acids Res. 41, 1151-1163.

Polikanov, Y. S., Melnikov, S. V., Soll, D. and Steitz, T. A. (2015). Structuralinsights into the role of rRNA modifications in protein synthesis and ribosomeassembly. Nat. Struct. Mol. Biol. 22, 342-344.

Schosserer, M., Minois, N., Angerer, T. B., Amring, M., Dellago, H., Harreither,E., Calle-Perez, A., Pircher, A., Gerstl, M. P., Pfeifenberger, S. et al. (2015).Methylation of ribosomal RNA by NSUN5 is a conserved mechanism modulatingorganismal lifespan. Nat. Commun. 6, 6158.

Sharma, S. and Lafontaine, D. L. J. (2015). ‘View From A Bridge’: a newperspective on eukaryotic rRNA base modification. Trends Biochem. Sci. 40,560-575.

Sharma, S., Watzinger, P., Kotter, P. and Entian, K.-D. (2013). Identification of anovel methyltransferase, Bmt2, responsible for the N-1-methyl-adenosine basemodification of 25S rRNA in Saccharomyces cerevisiae. Nucleic Acids Res. 41,5428-5443.

Shyu, Y. J., Suarez, C. D. and Hu, C.-D. (2008). Visualization of ternary complexesin living cells by using a BiFC-based FRET assay. Nat. Protoc. 3, 1693-1702.

Sloan, K. E., Bohnsack, M. T. andWatkins, N. J. (2013). The 5S RNP couples p53homeostasis to ribosome biogenesis and nucleolar stress. Cell Rep. 5, 237-247.

Steitz, T. A. and Moore, P. B. (2003). RNA, the first macromolecular catalyst: theribosome is a ribozyme. Trends Biochem. Sci. 28, 411-418.

Sulic, S., Panic, L., Barkic, M., Mercep, M., Uzelac, M. and Volarevic, S. (2005).Inactivation of S6 ribosomal protein gene in T lymphocytes activates a p53-dependent checkpoint response. Genes Dev. 19, 3070-3082.

Tilleray, V., Constantinou, C. and Clemens, M. J. (2006). Regulation of proteinsynthesis by inducible wild-type p53 in human lung carcinoma cells. FEBS Lett.580, 1766-1770.

Tuorto, F., Liebers, R., Musch, T., Schaefer, M., Hofmann, S., Kellner, S., Frye,M., Helm, M., Stoecklin, G. and Lyko, F. (2012). RNA cytosine methylation byDnmt2 andNSun2 promotes tRNA stability and protein synthesis.Nat. Struct. Mol.Biol. 19, 900-905.

Vousden, K. H. (2000). p53: death star. Cell 103, 691-694.White, J., Li, Z., Sardana, R., Bujnicki, J. M., Marcotte, E. M. and Johnson, A. W.

(2008). Bud23methylates G1575 of 18S rRNA and is required for efficient nuclearexport of pre-40S subunits. Mol. Cell Biol. 28, 3151-3161.

Yang, L., Song, T., Chen, L., Kabra, N., Zheng, H., Koomen, J., Seto, E. andChen, J. (2013). Regulation of SirT1-nucleomethylin binding by rRNA coordinatesribosome biogenesis with nutrient availability. Mol. Cell. Biol. 33, 3835-3848.

Zhang, Y. and Lu, H. (2009). Signaling to p53: ribosomal proteins find their way.Cancer Cell 16, 369-377.

Zhang, Y., Wolf, G. W., Bhat, K., Jin, A., Allio, T., Burkhart, W. A. and Xiong, Y.(2003). Ribosomal protein L11 negatively regulates oncoprotein MDM2 andmediates a p53-dependent ribosomal-stress checkpoint pathway. Mol. Cell. Biol.23, 8902-8912.

2393

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 2382-2393 doi:10.1242/jcs.183723

Journal

ofCe

llScience